Shengchao Liu

This paper introduces DuetRNA, a dual-frame flow matching model for RNA co-design that improves sequence-structure generation by jointly modeling base-and-sugar frames.

This paper introduces RigidSSL, a geometric pretraining framework for protein design and conformational modeling, improving designability and capturing realistic dynamics.

This paper introduces InertialGenome, a transformer-based framework for resolution-agnostic chromosome 3D structure reconstruction, showing improved performance across multiple resolutions and experimental validation.

GIMLET is a graph-text model that uses natural language instructions for zero-shot molecule property prediction, outperforming baselines and approaching supervised models.

The paper predicts drug classification from chemical structure using CNN and random forests, achieving better accuracy than transcriptomic-based methods.